“AI Can’t Solve the Crisis”: Christian Albrecht on Crisis Communication with AI

“AI Can’t Solve the Crisis”: Christian Albrecht on Crisis Communication with AI

How does artificial intelligence help identify communication crises earlier, evaluate information more quickly, and maintain the ability to act under time pressure? Christian Albrecht, Head of Communications at Nuremberg Airport, relies on AI-supported processes, a verified set of core facts, and clear limits on automation. In the CommTech Academy webinar on October 9, 2026, he used concrete real-world examples to demonstrate why it is not the technology itself, but rather the organization of crisis communication, that determines its effectiveness. AG CommTech spoke with him about how AI can serve as a crisis aid—and when humans absolutely must have the final say.

Christian Albrecht is a lecturer at the CommTech Academy. He will be leading a four-hour seminar there on November 9 on the topic of “Crisis Communication with AI.” Click here for details and to register.

AI cannot solve the crisis. But it can give us the crucial head start we need.

AG CommTech: Christian, when was the last time AI actually helped you in a critical situation?

Christian Albrecht: A few weeks ago, a storm swept through Nuremberg Airport. Trees were uprooted, security fences were damaged, and flight operations had to be temporarily suspended. Passengers wanted to know immediately how this would affect their trips. AI helped us create appropriate communication formats more quickly based on verified information. This is especially helpful when staff resources are limited.

AG CommTech: Where does AI have the greatest potential in crisis communication?

Christian Albrecht: It’s not just about writing the text. Three phases are crucial: the pre-crisis phase, the immediate response, and the follow-up. AI detects warning signs, supports situation assessments and reports, and subsequently assists with analysis. The rule of thumb is: First clarify processes, roles, and data quality; then select tools. If reporting chains aren’t working and no reliable information is available, even the most powerful AI won’t help.

Detect Crises Earlier with Social Listening

AG CommTech: How can communications departments identify crises earlier?

Christian Albrecht: Social listening allows us to monitor public discussions and noticeable changes. It’s also important to track how quickly topics gain traction. Systems like Brandwatch, Meltwater, Talkwalker, or NewsWhip provide relevant signals. Generative AI can use this data to create a concise briefing. Automated workflows can then send this briefing to the appropriate people, for example, via email or through Teams. This may free up some time.

AG CommTech: Have you ever found yourself lacking that kind of head start?

Christian Albrecht: The global grounding of the Boeing 737 MAX. At first, I saw it as an international aviation issue. Then journalists started asking: How many planes and passengers are affected in Nuremberg? I could have considered that local perspective earlier. An AI-powered analysis can help break down global developments to the level of one’s own organization and anticipate likely questions.

In the midst of an acute crisis: An overview of the situation and verified key facts

AG CommTech: During an acute crisis, rumors and information from many sources flood the communications team. How does AI help in this situation?

Christian Albrecht: Imagine a fire in a factory building. Images of smoke and speculation about an explosion are circulating on social media. Employees report fire trucks on the scene, journalists ask about injuries, and at the same time, the control center and plant management provide new information. AI can sort through these fragments, group related topics, flag open questions, and prepare a situation report. This helps ensure that rumors don’t become facts. Verifying and confirming the information remains our responsibility.

AG CommTech: How can this be used to quickly create consistent messages for different channels?

Christian Albrecht: The principle is simple: a verified core of facts, multiple outputs. We document what has been confirmed and where uncertainty still exists. From this, AI can draft an initial position statement, a spokesperson’s briefing, a Q&A, a website announcement, social media posts, and employee communications. All channels use the same foundation. Of course, final approval remains with a human. Linguistic quality is no substitute for facts.

AG CommTech: Your airport uses a chatbot. In a crisis, is it helpful or does it pose an additional risk?

Christian Albrecht: Both are possible. Our website is a central source of information. If there’s a current service disruption notice posted there, the chatbot must also be aware of that information. Otherwise, the website will report an incident while the chatbot says everything is fine. We’ve observed such inconsistencies elsewhere. That’s why a chatbot needs a controlled, up-to-date knowledge base and clear processes for crisis information. Otherwise, it only increases uncertainty. We’ve published details on this at kom.de.

AG CommTech: To what extent can the creation and distribution of crisis alerts be automated?

Christian Albrecht: Technically, many aspects can be integrated: evaluating signals from monitoring, creating a situation overview, and automatically forwarding it to the appropriate personnel. Draft texts, translations, and variations for different channels are also well-suited for automation. But automation is not an end in itself. AI can prepare a holding statement; the communications team decides whether its content is accurate, appropriate, and ready for publication. The closer you are to those affected and to publication, the more important human oversight becomes.

The closer to those affected and to publication, the more important human oversight becomes.

Christian Albrecht

Disinformation and AI-Generated Fakes

AG CommTech: AI is also being used to spread disinformation. How can we protect ourselves from fake images of crises?

Christian Albrecht: An airline was confronted with an AI-generated video purporting to show a plane crash. The media also picked up on it at first. The alleged incident was debunked through follow-up inquiries. Services like Google Lens, InVID-WeVerify, or provenance tools such as Content Credentials can help verify visual content. But no tool can replace independent verification. Especially when it comes to dramatic images, we must remain skeptical before reacting or sharing content.

Training, Boundaries, and Follow-Up

AG CommTech: What do you think of AI as a training partner for crisis situations?

Christian Albrecht: This is a particularly easy way to get started. A language model can play the role of an investigative journalist, ask critical questions, follow up on evasive answers, and ramp up the pressure. Afterward, there can be feedback on factual accuracy, empathy, and action orientation. You don’t need any special software for this. While such exercises are no substitute for a full-scale crisis response drill, they offer a straightforward way to practice your own ability to respond.

AG CommTech: For which tasks should AI never be allowed to make decisions on its own?

Christian Albrecht: When confirming facts, making strategic decisions, conducting risk assessments, and giving final approval. Statements regarding casualty figures, blame, responsibility, apologies, and expressions of sympathy are particularly sensitive. Here, we need human judgment and empathy. AI can prepare, structure, and provide suggestions, but it bears no responsibility. That is why companies must specify which systems are permitted, what data may be processed, and who reviews the results. This applies to the entire crisis management team, not just the press office.

AG CommTech: What should communications teams do better with AI in the wake of a crisis?

Christian Albrecht: The follow-up phase is often underestimated. Afterward, we should ask: What information came in, and when? Where did delays occur? What worked, and what didn’t? AI can organize minutes and other records, identify patterns, and assist with the evaluation. It is crucial that these insights are subsequently incorporated into processes, responsibilities, and drills. Every crisis is different. But we shouldn’t make the same mistakes twice.

More on this topic

• Interview with Timo Radzik: Why AI Needs Process Clarity
• AI Agents Need More Than Just Good Prompts
• Interview: What AI Agents Can Do in Corporate Communications
• All Best Practices from the CommTech Working Group

More on this topic

AI agents need more than just good prompts
What AI Agents Can Do in Corporate Communications
More about Christian Albrecht